Approximating gecko setae via direct laserlithography

Omar Tricinci1,8 , Eric V Eason2, Carlo Filippeschi1, Alessio Mondini1,Barbara Mazzolai1, Nicola M Pugno3,4,5 , Mark R Cutkosky6 ,Francesco Greco1,7,8 and Virgilio Mattoli1,8

1 Center for Micro-BioRobotics @SSSA, Istituto Italiano di Tecnologia, Viale Rinaldo Piaggio 34, 56025Pontedera, Italy2 Lockheed Martin Advanced Technology Center, Palo Alto, CA 94304, United States of America3 Laboratory of Bio-Inspired and Graphene Nanomechanics, Department of Civil, Environmental andMechanical Engineering, University of Trento, Via Mesiano 77, 38123 Trento, Italy4 School of Engineering and Materials Science, Queen Mary University of London, Mile End Road, E14NS London, United Kingdom5Ket Lab, Edoardo Amaldi Foundation, Italian Space Agency, Via del Politecnico snc, 00133 Rome, Italy6 Biomimetics and Dexterous Manipulation Laboratory, Center for Design Research, Department ofMechanical Engineering, Stanford University, Stanford, CA 94305 United States of America7Dept. of Life Science and Medical Bio-Science, Graduate School of Advanced Science and Engineering,Waseda University, 2-2 Wakamatsu-cho, 169-8480 Shinjuku-ku Tokyo, Japan

E-mail: omartricinci@gmail.com, francesco.greco@iit.it and virgilio.mattoli@iit.it

Received 22 September 2017, revised 14 November 2017Accepted for publication 30 November 2017Published 29 May 2018

AbstractThe biomimetic replication of dry adhesion present in the gecko’s foot has attracted great interestin recent years. All the microfabrication techniques used so far were not able to faithfullyreproduce the hierarchical and complex three-dimensional geometry of the gecko’s setae, withfeatures at the micro- and nano-scale, thus reducing the effectiveness that such conformalmorphology could provide. By means of direct laser lithography we fabricated artificial hairs thatfaithfully reproduce the natural model. This technique allows the fabrication of three-dimensional microstructures with outstanding results in terms of reproducibility and resolution atthe micro- and nano-scale. It was possible to get very close to the morphology of the naturalgecko setae, especially concerning the hierarchical shape. We designed several morphologies forthe setae and studied the effects in terms of adhesion and friction performances compared to thenatural counterpart, showing the interplay between morphology, dimensional scaling andmaterials. Direct laser lithography promises great applications in the biomimetics field, pavingthe way to the implementation of the concept of hierarchical bioinspired dry adhesives.

Supplementary material for this article is available online

Keywords: biomimetics, two-photon lithography, gecko, dry adhesion, biomimetic hierarchicalmaterials

(Some figures may appear in colour only in the online journal)

1. Introduction

The fabrication of smart surfaces based on biomimetic prin-ciples has attracted great interest in recent years [1–7]. In

particular, several studies have tried to reproduce theremarkable capability of dry adhesion that is present in thegecko’s foot [8–22]. Geckos can easily climb surfaces withvarious degrees of roughness, supporting their relatively highweight thanks to the hierarchical conformal morphology ofthe pads at the micro- and nano-scale. The gecko’s adhesive

Smart Materials and Structures

Smart Mater. Struct. 27 (2018) 075009 (9pp) https://doi.org/10.1088/1361-665X/aa9e5f

8 Authors to whom any correspondence should be addressed.

0964-1726/18/075009+09$33.00 © 2018 IOP Publishing Ltd Printed in the UK1

system is directional and controllable: it provides immediateand strong adhesion when loaded in a particular direction andit requires very low force for attachment and detachment. Inaddition, it is ‘self cleaning’ and does not attract dirt; hence itcan be reused many times without degrading its performance[23]. These characteristics would be desirable for climbingand grasping applications in robotics.

Although numerous dry adhesives have been createdusing lithographic and micromachining techniques, none ofthese fully realizes the capabilities of the gecko’s adhesivesystem. One important reason may be that it is difficult toreproduce the complex three-dimensional morphology of thegecko’s lamellae, setae and spatula with features sizes below1 μm. Instead, lithographic processes produce arrays of pil-lars, wedges, stalks or ‘mushroom cap’ features that produceuseful levels of van der Waals adhesion when pressed againstsmooth and clean surfaces, but not on surfaces with highlevels of roughness [24]. For better conformability, a numberof investigations have produced hierarchical features, how-ever these usually come at the expense of decreased real areaof contact, and hence reduced adhesion, on smooth surfaces[25]. Here, by means of a 3D laser lithography technique,complex artificial hairs inspired by the gecko pad were fab-ricated, with a geometry that approximates the biologicalmodel. Direct 3D laser lithography has already demonstratedits potential in the realization of biomimetic surfaces pat-terned with three-dimensional features at sub-micrometricresolution [26–29]. The limiting factor of the proposedtechnique is related to the relatively large fabrication timeeven if advanced 3D laser lithography systems could over-come such issues (Galvo mode, Nanoscribe GmbH). In thiswork we also validated the adhesive and friction perfor-mances of the artificial setae, investigating different designsand showing the crucial interplay between morphology,dimensional scaling and materials.

The study started from a detailed design of the setae,based on the anatomy of gecko pads, which has been thor-oughly investigated [1, 23]. The gecko foot skin is composedof a multilevel structure of lamellae, setae, branches andspatulae primarily composed of β-keratin, whose elasticmodulus is 1.5 GPa [30]. The setae have a length in the rangeof 30–130 μm and a diameter of about 5–10 μm, while thespatulae—the end portions which ramify from the branches—have final tips 0.5 μm long, 0.2–0.3 μm wide and 10 nm thick(figure 1(a)). Each segment, thanks to its independent defor-mation, easily conforms and comes in close contact to solidsurfaces at different levels of roughness, from the mm to thenm scale, providing a huge number of short range interactionsmainly governed by van der Waals forces [30, 31]. In the caseof Gekko gecko, each of the 5000 setae mm−2 can ensure anaverage adhesion force of about 20 μN, although the max-imum adhesion can reach a value tenfold higher, dependingon the preload [32].

The high aspect ratio of the stalk, the hierarchicalmorphology and the micro- and nano-scale features of thebranches make the replication of such a complex design

extremely challenging with most microfabrication techniques.The artificial setae constructed in this work are composed of(I) a stalk, (II) two levels of branches and (III) the finalspatulae (figure 1(b)). Different designs were obtained bytuning the angles between the branches. We focused ourinvestigation on the friction force and adhesive normal forceof the artificial setae, which play a key role in the so-calledadhesive friction or parallel adhesion that is exploited by thegeckos [33].

Finally, it seems worthy to remark the goal of ourinvestigation since the fabrication of dry adhesives could alsobe obtained by means of a range of inorganic materials thatthe scientist can nowadays choose according to practicalissues. Nature, on the other hand has developed a peculiarshape for the setae due to materials constraints.

2. Experimental methods

2.1. Microfabrication procedure

The setae were fabricated in IP-DiLL photoresist (NanoscribeGmbH) by means of a Nanoscribe system (NanoscribeGmbH). The elastic modulus of the photoresist is 1–3 GPa,according to the laser power and the writing speed [36],comparable to the natural counterpart, thus being a goodcandidate material for the artificial structures. A square siliconsample of side 1 mm was glued on a glass slide; the resist,poured on the silicon substrate, was exposed to a laser beamat a center wavelength of 780 nm, using a writing speed of100 μm s−1 with a power of 5.6 mW (Calman laser source).The sample was developed for 20 min in SU-8 Developer(MicroChem Corp) and rinsed in IPA and deionized water.

2.2. Design of artificial setae

The fabrication procedure required the design of two differentmodels. The first model tried to reproduce the roughmorphology of the gecko setae and it was used for the cali-bration of the main fabrication parameters such as the laserpower and the writing speed (figure S1 in SI is availableonline at stacks.iop.org/SMS/27/075009/mmedia). Thesecond model was designed, in several slightly differentvariant morphologies, in order to reproduce the effectiveshape of the natural setae with a higher level of detail. Thestructures fabricated according to this type of model wereused for the adhesion and friction tests. In both cases the tipshave been modeled as paths of coordinates in Matlab. Thesorting strategy is based on a breadth-first algorithm, in whichall adjacent edges at the same depth from the base are writtenbefore moving to the next level of edges that are further fromthe base. The programmed stalk has a nominal length of60 μm and a square section with nominal sides of 3.4 μm; itforms an angle of 30° relative to the substrate. The branchesgrow of about 20 μm along the direction perpendicular to thesubstrate. There are 16 rectangular final tips with sides of1 μm and 1.2 μm (figures 1(c)–(e)). Once the lithographic

2

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

Figure 1. (a) Scanning electron micrographs reporting the hierarchical adhesive structures of Gekko gecko at different magnifications; ST:seta, SP: spatula, BR: branch (adapted from Gao et al 2005 [34], copyright 2005, with permission from Elsevier). (b) Simulation of theartificial programmed setae showing the first level of branching, the second level and the final spatulae (green, red and blue bar respectively).(c) First prototype of setae and details of branches (d) and spatulae ((e) adapted from Tricinci et al 2017 [35] © Springer InternationalPublishing AG, with permission of Springer). (f) Example of a definitive design and details of branches in (g). (h) View from the top of thedefinitive design, showing the details of the spatulae.

3

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

parameters have been defined, they have been used for thefabrication of the setae for the experimental tests. A Matlabprogram was written for the implementation of differentgeometrical parameters of the setae which may affect theadhesion properties. In the definitive model the stalk has anominal length of 60 μm and it has a square section withnominal sides of 2.8 μm and it forms an angle of 30° relativeto the substrate. There are three levels of branches that endwith 128 tips in total, each of 300 nm in thickness and 1 μm inlength (figures 1(f)–(h)).

2.3. Friction and adhesion tests

The adhesion and friction tests were carried out under anoptical microscope (KH-7700 digital microscope, Hirox). Themicroforce sensor (FemtoTools GmbH) was mounted on amicromanipulator, while the silicon sample with the artificialgecko setae was fixed on the tip of a nanomanipulator(Kleindiek Nanotechnik GmbH), as shown in figure 2(a). Thesurface of the sensing component was silicon. The sample andthe sensor were properly aligned under the microscope(figure 2(b)). The cycles of the tests were controlled by meansof dedicated software code written with Visual Basic(Microsoft Corp.) and data were acquired at a frequency of100 Hz. Since the microforce sensor operates only along the

direction of its axis, and therefore it was not possible tomeasure at the same time the perpendicular preload and theeffective friction force, a calibration procedure for the esti-mation of the perpendicular preload as a function of thebending of the setae was required.

For the evaluation of the friction on silicon, the testcycles consisted of: (i) approaching and perpendicular pre-loading, (ii) parallel displacement of about 5 μm, (iii) slidingof the setae along the surface of the microforce sensor duringwhich the parallel adhesive force is measured, and (iv) finalperpendicular detachment (figure 2(c)). The measurement ofthe friction force was carried out in a quasi-static condition, ata speed of 35 nm s−1.

The test for the perpendicular adhesion of the artificialsetae was carried out on two different target surface materialswith different stiffness: silicon, that is hard, and poly-dimethylsiloxane (PDMS), that is softer. The PDMS surfacewas obtained by dip coating the silicon sensor. The sampleswere moved against the sensor with a speed of 5 μm s−1, untilreaching a preset load that varied depending on the geometryand the tested surface. Then a parallel slide of 10 μm wasimposed to provide the rearrangement of the tips. Finally thesamples were detached (in the perpendicular direction) with aspeed of 5 μm s−1.

Figure 2. (a) Scheme of the experimental setup for the estimation of the friction and adhesion force. (b) Optical image of the artificialmicrofabricated setae aligned with the sensor cantilever. (c) Phases of the cycle of friction tests: (i) approach and perpendicular preload;(ii) parallel preload; (iii) parallel quasi-static sliding; (iv) detachment.

4

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

3. Results and discussion

Ten different designs of the setae obtained by changing theangles between the branches (labelled from A to J,figure 3(a)) were tested, assessing the effect of density/

proximity of spatulae on the friction and normal adhesiveforces. The results of the microfabrication procedure areillustrated in figures S2–3 in SI.

In figure 3(b) the typical behaviour of the parallel fric-tional force as a function of time is depicted.

Figure 3. (a) Scheme of the setae highlighting angles changed in the designs A-J and corresponding table summarizing different designs.(b) Friction force of a seta (design C) pulled parallel to the surface of the silicon sensor, plotted as a function of time, in a typical experiment.The four phases of the friction test are labelled above the graph: (i) approach and perpendicular preload; (ii) parallel preload; (iii) parallelquasi-static sliding; (iv) detachment. (c) Friction force of all the setae designs at 5 μN of perpendicular preload; the dotted line represents theperformance of the control sample (1.2 μN). (d) Friction force normalized to the actual area of each artificial setae (5 μN of perpendicularpreload); the dotted line represents the performance of the natural setae (0.1 N mm−2) [32]. (e) Control sample with hemispherical tip.

5

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

A prominent factor affecting the friction is the preloadimposed on the setae. For this reason, emulating the beha-viour of the gecko during its walk, both perpendicular andparallel preload forces were applied [32]. The perpendicularpreload force was applied perpendicularly against the sub-strate while the parallel preload was applied by means of asmall parallel displacement; this increases the number of thecontact points as the spatulae lay over and conform to thesubstrate.

Experimental data of the friction force vs perpendicularpreload (figure S4 in SI) are compatible with a linear rela-tionship that would be expected in frictional phenomena.

Designs I, B and H ensure good friction (maximum slopeof the curve in figure S4 in SI), in absolute terms (figure 3(c)).To confirm the effectiveness of the hierarchical configuration,we carried out the same experiment on a control sampleconsisting of a single cylindrical pillar composed by 4 seg-ments with the same inclination and section area of thecorresponding level of the tested samples and with a semi-spherical tip (figure S5 in SI and figure 3(e)). Such shape wasdesigned in order to have the same contact area of the sam-ples, ensuring a permanent contact that was very hard toachieve with a flat control surface. Comparing the results ofthe friction (at 5 μN of perpendicular preload) of the controlwith those of the hierarchical samples, we found that thecontrol sample was weaker, demonstrating that the hier-archical configuration was more effective; in our case, theartificial setae produced forces 6–9 times higher than thecontrol.

To compare the performances of the artificial setae withthe natural ones, it is necessary to normalize their contactarea. In particular, the range of interest is around 5 μN ofperpendicular preload, since this is the typical value observedin gecko setae, which, in this condition, can provide anaverage force of 20 μN, that is a normalized force of0.1 N mm−2 [32]. The area covered by the artificial spatulaeranges between about 35% and 80% (depending on the spe-cific design considered) of the area of the natural ones, due tothe different angles between the branches. Broader anglesbetween the branches (as in cases H and I) produce a largearea of contact and friction force, as shown in figure 3(c), buta lower effectiveness when normalized by area (figure 3(d)).Conversely, a configuration with closer spatulae (as in casesA and B) ensures high values of force per unit area, com-parable with the natural setae since the spatulae have greaterdensity. The best design solution should have an increase ofthe density of the artificial spatulae while avoiding self-sticking or bunching: this can be obtained by reducing thedimensions of the spatula, which is exactly what happens inthe natural gecko setae.

Although the friction force of artificial setae is compar-able to the natural counterpart, their conformability to everytype of surface is intrinsically limited, since they do notinclude morphological features down to the tens of nan-ometers, which are instead present in gecko setae [32]. Wenote also that the surface of the sensor used for the tests playsa role in the friction; although it is flat it is not smooth at the

nanometer scale but is characterized by a roughness withfeatures comparable to the natural spatulae (figure S6 in SI).For all these reasons it is expected that artificial setae willexert lower forces than gecko setae.

In order to investigate this aspect we carried out thesecond experiment in which we tested the normal adhesion ofthe artificial setae. In this case, it has to be considered that inthe artificial setae the spatulae are a series of lines obtained byextruding a solid ellipsoid voxel, so that the terminal part ofthe tips involved in the contact process can be approximatedto a hemisphere. For this reason the interaction of each featurein the spatulae with a flat surface can be modelled as that of asphere of radius R. In this condition van der Waals interactionenergy WVDW between a sphere and a flat surface is [37]:

= - ( )W , 1VDWAR

D6

where D is the distance between the body and the surface(D=R) and A is the Hamaker constant in the range of10−20

—10−19J (A=π2Cρ1ρ

2 ,where ρ

1and ρ

2are the number

of atoms per unit volume in the two interacting bodies) [37].The related van der Waals normal force is:

= - ( )F . 2VDWAR

D6 2

The van der Waals normal force between the setae and aflat surface is calculated as:

= - ( )F , 3VDWARn n

D6t s2

where nt and ns are the number of tips in the spatulae and thenumber of spatulae in the setae, respectively 2 and 64; D is anatomic gap distance of 0.3 nm. The value of R was chosen tobe 150 nm which represents the maximum radius of theellipsoid along the longitudinal direction, thus providing asuperior limit of confidence. According to this model, we canestimate the normal adhesion force for this type of artificialsetae in the range from 0.36 to 3.6 μN.

We tested the normal adhesion of the artificial setae(designs with the closest (A) and broadest (J) branches) to twodifferent target surface materials: silicon and poly-dimethylsiloxane (PDMS). PDMS was chosen for its relativesoftness, thus trying to find a material surface able to help thesetae to better conform. In the case of design A each sampleconsisted of six setae, while in the case of the design J eachsample had 4 setae. During this step, if adhesion occurred, thenormal force flipped sign from positive to negative becausethe attached setae pulled the sensor while being retracted.Figure 4 summarizes the results obtained in normal adhesiontests. A bare Si surface was first considered as the targetsurface. The combination of design A against the Si surfacedid not produce any measurable adhesion (figure 4(a)) sincethe closely arranged tips did not properly interact withthe rigid silicon surface. A similar result was observed in thecase of design J against Si, even though the tips had anapparently good configuration (figure 4(b)). Several peakswere observed during the force increase/decrease steps,underlining how the tips had the time to adjust their con-formation over the surface of the sensor. A second set ofexperiments considered PDMS as the target surface, obtained

6

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

by dip coating the silicon sensor. The combination of designA against PDMS produced measureable adhesion(figure 4(d)) since the high compliance of the PDMS wasenough to provide a proper interaction with the tips. Theadhesion force recorded was 1.88±0.11 μN, with a meanadhesion force of 0.31 μN per seta. This small value is notwithin the interval predicted by the mathematical model. Thereason for such an apparently weaker adhesion could be thatnot all the tips were in contact with the surface of the sensor,or that each tip was not fully aligned with the surface.Improved adhesion was observed in the case of design Jagainst PDMS (figure 4(e)). Rearrangement of tips is againobserved as a short sawtooth profile in the graphic. The

recorded adhesion force was 1.82±0.15 μN, which means amean single seta force of 0.46 μN, inside the predicted valuerange and considerably higher than design A. The result ispromising since, as already evidenced for design A, not all ofthe setae may have made perfect simultaneous contact withthe surface of the sensor. In contrast, the control sampleshowed no measureable adhesion on either the Si or PDMSsurfaces (figures 4(c), (f)). Although the obtained values arefar from the performance of the natural setae, we were able tomeasure, for the best configuration, around 25 g cm−2 ofadhesion that is not only due to the target material (PDMS)adhesiveness. In fact, from these tests we can deduce thatbroader artificial spatulae are able to exert improved adhesion

Figure 4. Normal adhesion tests of artificial setae with different designs against target surfaces: (a) Normal force on Si of design A(perpendicular preload 20 μN) and in (b) design J (perpendicular preload 22 μN) and in (c) the control (perpendicular preload 4.5 μN);(d) normal force on PDMS of design A (perpendicular preload 30 μN) and in (e) design J (perpendicular preload 13 μN) and in (f) the control(perpendicular preload 8 μN).

7

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

forces with respect to narrower designs since they can betterconform to the surface; nevertheless a higher density ofspatulae per surface area would be required to obtain aneffective force comparable to the gecko setae. These tworequirements must be balanced according to the minimumsize of features that can be fabricated with the proposedtechnique.

Additionally, it is important to remark that, while thepresent study mainly focused on structural features as a meansto tune adhesion of artificial setae, the material composition(both of the artificial setae and target surface) plays a role in theadhesion. Focusing on this aspect, the effective elastic modulusrepresents a critical requirement: although β-keratin and IP-DiLL have comparable elastic modulus, the effective modulusis lower in the natural setae since they have a slimmer shaperespect to the artificial ones. Our structures are still slightly toorigid to properly conform even on a flat surface. A soft materialplaced at least on the tips would be of some help.

This consideration suggests the possibility of fabricatingthe artificial setae with different structural materials or per-forming a surface functionalization to optimize their smartadhesive behaviour for specific target surfaces.

4. Conclusion

In conclusion, gecko foot-like microstructures fabricated bymeans of direct laser lithography have been presented. Thesebiomimetic structures reproduced with unprecedented faith-fulness the structural microscale features of the animal model.The artificial setae were fabricated with different designs toinvestigate how their morphology influences their adhesiveperformance. Adhesion tests of the artificial setae developedhere provided meaningful insight into their structure/propertyrelationship. We investigated the friction force and adhesivenormal force of the artificial setae, which are at the basis ofthe frictional adhesion in the geckos’ toe.

In particular, friction forces similar to the natural modelwere obtained using optimized designs. On the other hand,the adhesive force of the artificial setae is still weaker than inthe geckos’ setae, due to the mechanical properties of thestructures. To achieve higher forces and reach the goal ofemulating the natural model, the interplay of morphology,dimensional scaling and materials involved should be con-sidered. This work represents a first step in the direction ofartificial biomimetic microstructures, with nanometric fea-tures, that closely mimic the morphology of the natural geckosetae. Selecting new structural materials would be in principlepossible to enhance and optimize the adhesion on virtuallyevery kind of target surface.

Acknowledgments

N.M.P. is supported by the European Commission H2020under the Graphene Flagship Core 2 N. 785219 (WP14‘Polymer Composites’), and FET Proactive ‘Neurofibre’Grant No. 732344.

FG acknowledges financial support from Top GlobalUniversity Project (Unit for Energy and Nanomaterials) atWaseda University from MEXT Japan. EVE and MRC weresupported in part by the DARPA Zman project.

ORCID iDs

Omar Tricinci https://orcid.org/0000-0002-8402-0628Nicola M Pugno https://orcid.org/0000-0003-2136-2396Mark R Cutkosky https://orcid.org/0000-0003-4730-0900Virgilio Mattoli https://orcid.org/0000-0002-4715-8353

References

[1] Bhushan B 2009 Biomimetics: lessons from nature–anoverview Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 3671445–86

[2] Koch K, Bhushan B and Barthlott W 2009 Multifunctionalsurface structures of plants: an inspiration for biomimeticsProg. Mater. Sci. 54 137–78

[3] Kinoshita S, Yoshioka S and Kawagoe K 2002 Mechanisms ofstructural colour in the morpho butterfly: cooperation ofregularity and irregularity in an iridescent scale Proc. R. Soc.Lond. B Biol. Sci. 269 1417–21

[4] Dean B and Bhushan B 2010 Shark-skin surfaces for fluid-dragreduction in turbulent flow: a review Philos. Trans. R. Soc.Math. Phys. Eng. Sci. 368 4775–806

[5] Stavenga D G, Foletti S, Palasantzas G and Arikawa K 2006Light on the moth-eye corneal nipple array of butterfliesProc. R. Soc. B Biol. Sci. 273 661–7

[6] Jagota A and Hui C Y 2011 Adhesion, friction, and complianceof bio-mimetic and bio-inspired structured interfaces Mater.Sci. Eng. R Rep. 72 253–92

[7] Hou Y, Yu M, Chen X, Wang Z and Yao S 2015 Recurrentfilmwise and dropwise condensation on a beetle mimeticsurface ACS Nano 9 71–81

[8] Kamperman M, Kroner E, del Campo A, McMeeking R M andArzt E 2010 Functional adhesive surfaces with ‘gecko’effect: the concept of contact splitting Adv. Eng. Mater. 12335–48

[9] Autumn K, Niewiarowski P H and Puthoff J B 2014 Geckoadhesion as a model system for integrative biology,interdisciplinary science, and bioinspired engineering Annu.Rev. Ecol. Evol. Syst. 45 445–70

[10] Geim A K, Dubonos S V, Grigorieva I V, Novoselov K S,Zhukov A A and Shapoval S Y 2003 Microfabricatedadhesive mimicking gecko foot-hair Nat. Mater. 2 461–3

[11] Northen M T and Turner K L 2005 A batch fabricatedbiomimetic dry adhesive Nanotechnology 16 1159–66

[12] Lee H, Lee B P and Messersmith P B 2007 A reversible wet/dry adhesive inspired by mussels and geckos Nature 448338–41

[13] Lu G, Hong W, Tong L, Bai H, Wei Y and Shi G 2008 Dryingenhanced adhesion of polythiophene nanotubule arrays onsmooth surfaces ACS Nano 2 2342–8

[14] Murphy M P, Kim S and Sitti M 2009 Enhanced adhesion bygecko-inspired hierarchical fibrillar adhesives ACS Appl.Mater. Interfaces 1 849–55

[15] Lee J, Bush B, Maboudian R and Fearing R S 2009 Gecko-inspired combined lamellar and nanofibrillar array foradhesion on nonplanar surface Langmuir 25 12449–53

[16] Greiner C, Arzt E and del Campo A 2009 Hierarchical gecko-like adhesives Adv. Mater. 21 479–82

8

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al

[17] Kim T, Jeong H E, Suh K Y and Lee H H 2009 Stoopednanohairs: geometry-controllable, unidirectional, reversible,and robust gecko-like dry adhesive Adv. Mater. 21 2276–81

[18] Kim T, Pang C and Suh K Y 2009 Shape-tunable polymernanofibrillar structures by oblique electron beam irradiationLangmuir 25 8879–82

[19] Mengüç Y, Yang S Y, Kim S, Rogers J A and Sitti M 2012Gecko-inspired controllable adhesive structures applied tomicromanipulation Adv. Funct. Mater. 22 1246–54

[20] Jeong H E, Kwak M K and Suh K Y 2010 Stretchable,adhesion-tunable dry adhesive by surface wrinklingLangmuir 26 2223–6

[21] Hawkes E W, Eason E V, Asbeck A T and Cutkosky M R 2013The Gecko’s toe: scaling directional adhesives for climbingapplications IEEEASME Trans. Mechatron. 18 518–26

[22] King D R, Bartlett M D, Gilman C A, Irschick D J andCrosby A J 2014 Creating gecko-like adhesives for ‘realworld’ surfaces Adv. Mater. 26 4345–51

[23] Autumn K and Gravish N 2008 Gecko adhesion: evolutionarynanotechnology Phil. Trans. R. Soc. A 366 1575–90

[24] Hawkes E W, Eason E V, Christensen D L and Cutkosky M R2014 Human climbing with efficiently scaled gecko-inspireddry adhesives J. R. Soc. Interface 12 20140675

[25] Kwak M K, Pang C, Jeong H-E, Kim H-N, Yoon H,Jung H-S and Suh K-Y 2011 Towards the next level ofbioinspired dry adhesives: new designs and applicationsAdv. Funct. Mater. 21 3606–16

[26] Marino A, Filippeschi C, Mattoli V, Mazzolai B and Ciofani G2015 Biomimicry at the nanoscale: current research andperspectives of two-photon polymerization Nanoscale 72841–50

[27] Tricinci O, Terencio T, Mazzolai B, Pugno N M, Greco F andMattoli V 2015 3D micropatterned surface inspired bysalvinia molesta via direct laser lithography ACS Appl.Mater. Interfaces 7 25560–7

[28] Bernardeschi I, Tricinci O, Mattoli V, Filippeschi C,Mazzolai B and Beccai L 2016 Three-dimensional softmaterial micropatterning via direct laser lithography offlexible molds ACS Appl. Mater. Interfaces 8 25019–23

[29] Röhrig M, Thiel M, Worgull M and Hölscher H 2012 3D directlaser writing of nano- and microstructured hierarchicalgecko-mimicking surfaces Small 8 3009–15

[30] Autumn K, Majidi C, Groff R E, Dittmore A and Fearing R2006 Effective elastic modulus of isolated gecko setal arraysJ. Exp. Biol. 209 3558–68

[31] Autumn K, Sitti M, Liang Y A, Peattie A M, Hansen W R,Sponberg S, Kenny T W, Fearing R, Israelachvili J N andFull R J 2002 Evidence for van der waals adhesion in geckosetae Proc. Natl Acad. Sci. 99 12252–6

[32] Autumn K, Liang Y A, Hsieh S T, Zesch W, Chan W P,Kenny T W, Fearing R and Full R J 2000 Adhesive force ofa single gecko foot-hair Nature 405 681–5

[33] Autumn K, Dittmore A, Santos D, Spenko M and Cutkosky M2006 Frictional adhesion: a new angle on gecko attachmentJ. Exp. Biol. 209 3569–79

[34] Gao H, Wang X, Yao H, Gorb S and Arzt E 2005 Mechanics ofhierarchical adhesion structures of geckos Mech. Mater. 37275–85

[35] Tricinci O, Eason E V, Filippeschi C, Mondini A, Mazzolai B,Pugno N M, Cutkosky M R, Greco F and Mattoli V 2017 Dryadhesion of artificial gecko setae fabricated via direct laserlithography Biomimetic and Biohybrid Systems; LectureNotes in Computer Science (Cham: Springer) pp 631–6

[36] Lemma E D, Rizzi F, Dattoma T, Spagnolo B, Sileo L,Qualtieri A, Vittorio M D and Pisanello F 2017 Mechanicalproperties tunability of three-dimensional polymeric structures intwo-photon lithography IEEE Trans. Nanotechnol. 16 23–31

[37] Israelachvili J N 2011 Intermolecular and Surface Forces:Revised 2rd edn Academic Press Waltham (Waltham, MA:Academic Press)

9

Smart Mater. Struct. 27 (2018) 075009 O Tricinci et al